DD
MM
YYYY

PAGES

DD
MM
YYYY

spot_img

PAGES

Home Blog Page 1643

DeepSeek 2025 Uptrend Updates and Its Open Source Model Could Accelerate AI Autonomy

0

DeepSeek, a Chinese AI startup, has been making waves with recent releases, but there’s no confirmed information about a brand-new AI model dropping as of April 30, 2025. The latest significant releases from DeepSeek include the DeepSeek-V3-0324 model, an upgrade to its V3 large language model, launched on March 24, 2025, and the R1 reasoning model, released in January 2025.

There’s also buzz around a potential R2 model, speculated to be the successor to R1, with Reuters reporting in March 2025 that DeepSeek was accelerating its launch, possibly targeting April 2025. However, no official confirmation or specific release details for R2 have surfaced in the available data.

This model, released via Hugging Face, boasts improved reasoning and coding capabilities over its December 2024 V3 predecessor. It’s a Mixture-of-Experts (MoE) model with 671 billion parameters, trained on 14.8 trillion tokens for about $5.6 million—significantly cheaper than competitors like OpenAI’s GPT-4. It’s open-source under the MIT License and competes with models like GPT-4o and Anthropic’s Claude 3.5 Sonnet.

A reasoning model also with 671 billion parameters, R1 matches or outperforms OpenAI’s o1 on benchmarks, particularly in math and coding. It’s open-source, cost-efficient (trained on Nvidia H800 GPUs), and sparked a tech stock sell-off due to its disruptive potential. Its chatbot app topped the U.S. iOS App Store, surpassing ChatGPT. Reports suggest DeepSeek is rushing to release R2, initially planned for early May 2025, to capitalize on R1’s success. It’s expected to enhance coding and multilingual reasoning, but DeepSeek has remained silent on specifics.

DeepSeek, collaborating with Tsinghua University, introduced a technique combining generative reward modeling (GRM) and self-principled critique tuning (SPCT). This aims to boost LLM performance, with plans to open-source the resulting DeepSeek-GRM models, though no release date is confirmed. The lack of concrete evidence for a new model beyond these suggests you might be referring to the V3-0324, R1, or the anticipated R2. DeepSeek’s rapid pace—releasing models like V3 in December 2024, R1 in January 2025, and V3-0324 in March 2025—shows they’re iterating fast, challenging U.S. giants like OpenAI with cost-effective, open-source alternatives.

Their approach, using techniques like MoE and optimization on less powerful chips, has rattled the industry, with some calling it “AI’s Sputnik moment.” The 2025 releases from DeepSeek, notably the DeepSeek-V3-0324 and R1 models, and the anticipated R2, have far-reaching implications for the AI industry, global tech competition, economic dynamics, and national security.

DeepSeek’s models, trained at a fraction of the cost of Western counterparts (e.g., V3 at $5.6 million vs. GPT-4’s estimated $100 million), challenge the assumption that massive computational resources are necessary for cutting-edge AI. This has several implications. The R1 model’s performance, rivaling OpenAI’s o1 at 4% of the cost, signals that large language models (LLMs) are becoming commoditized. This could erode the value of proprietary models, forcing companies like OpenAI to cut prices or shift to mass-market strategies.

By open-sourcing models like R1 and V3-0324 under the MIT License, DeepSeek enables smaller companies, startups, and developers in resource-constrained regions to build on its architecture. This democratizes AI innovation, potentially leading to a surge in specialized applications. DeepSeek’s use of techniques like Mixture-of-Experts (MoE), mixed-precision arithmetic, and optimized Nvidia H800 GPUs shows that software and hardware efficiency can rival brute-force scaling. This may push competitors to adopt similar approaches, accelerating cost declines (already down 80% annually pre-DeepSeek).

DeepSeek’s ability to match or surpass models like GPT-4o and o1, despite U.S. chip export controls, questions whether American firms can maintain their lead. President Trump called it a “wake-up call” for U.S. industries, highlighting concerns about competitiveness. U.S. sanctions on advanced chips (e.g., H100/A100) have pushed DeepSeek to innovate with less powerful H800 GPUs and techniques like PTX programming and Native Sparse Attention (NSA). This resilience suggests export controls may not halt Chinese progress, potentially isolating U.S. tech from China’s market.

DeepSeek’s success has spurred U.S. policy responses, including proposed bans on its app and restrictions on cloud providers offering its models. A bipartisan bill and congressional reports allege DeepSeek harvests U.S. data and uses banned Nvidia chips, raising national security concerns. These moves could escalate tech decoupling. DeepSeek’s releases triggered significant market reactions, with a $1 trillion-plus sell-off in global equities, including a record $589 billion single-day loss for Nvidia.

DeepSeek’s efficiency gains challenge the rationale for massive AI infrastructure spending (e.g., $371 billion by hyperscalers in 2025). Investors are questioning the viability of huge funding rounds for foundation model developers like OpenAI, which raised over $30 billion. As AI costs plummet, usage is expected to surge (Jevons paradox), shifting value from model training to inference and application-specific tasks. This could boost demand for custom chips (XPUs) and benefit AI adopters across industries.

In China, DeepSeek’s models are embedded in sectors like automotive e.g., Geely’s AI-powered cars, smartphones (e.g., Huawei’s Xiaoyi), and government services, reflecting Beijing’s push for AI-driven economic growth. This rapid adoption could give Chinese firms a competitive edge globally, especially in electric vehicles. U.S. lawmakers and the House Select Committee on China claim DeepSeek’s chatbot, hosted on Chinese servers, could harvest sensitive U.S. user data, acting as “AI spyware.” Weak data safeguards and alleged links to a banned Chinese telecom company amplify these concerns.

Freely downloadable models like R1 could harbor censorship controls or vulnerabilities, posing risks to global AI infrastructure if widely adopted. Distillation techniques, which DeepSeek uses to compress models, may perpetuate privacy issues from training data, now outside U.S. jurisdiction. DeepSeek’s efficiency could inspire U.S. firms to build compact, high-performing AI for military use, enhancing tools for the Pentagon. However, its open-source nature raises fears of adversaries accessing powerful AI.

By sharing code repositories and algorithms like NSA, DeepSeek fosters a collaborative AI ecosystem, potentially accelerating innovation and making models more transparent and trustworthy. R1 has already spawned thousands of derivative models. The success of R1, a reasoning model using chain-of-thought (CoT) techniques, has shifted industry attention to models that solve problems step-by-step, requiring more inference compute. This could drive investment in reasoning-intensive applications.

DeepSeek’s energy-efficient models, using less memory and compute, offer a path to greener AI, addressing concerns about AI’s high carbon footprint. Open-source models could democratize AI, enabling broader societal benefits, but also raise risks of misuse (e.g., biased outputs or falsehoods) due to less oversight compared to proprietary models.

DeepSeek’s collaboration with Tsinghua University on self-improving models (e.g., self-principled critique tuning, SPCT) and plans to open-source DeepSeek-GRM models suggest a focus on autonomous, efficient AI. This could. Self-improving models may reduce reliance on human fine-tuning, lowering costs further and enabling faster iteration.

Posts on X speculate that enhancements like memory, long-context handling, or agentic capabilities in R2 could “put U.S. frontier labs in shambles.” While unverified, this reflects sentiment that DeepSeek’s trajectory threatens Western dominance. Self-improving AI and open-source distribution amplify concerns about control, safety, and unintended consequences, necessitating robust governance frameworks.

DeepSeek’s 2025 releases have upended AI assumptions, proving that cost-efficient, open-source models can rival proprietary giants. This shift lowers barriers to AI innovation, intensifies U.S.-China competition, and reshapes economic priorities toward applications and inference. However, it also raises critical security, privacy, and ethical challenges, particularly given DeepSeek’s Chinese origins and open-source approach.

The anticipated R2 release, potentially imminent, could amplify these trends, with AI analysts suggesting it may push boundaries in multilingual reasoning and efficiency. For stakeholders, the challenge is balancing the benefits of accessible AI with the risks of unchecked proliferation and geopolitical fallout.

Alphabet CEO Sundar Pichai Warns DOJ’s Remedies Will Result in ‘De Facto Divestiture’, Stripping Google of Its Core Business

0

In a striking turn at the federal courthouse in Washington, Alphabet CEO Sundar Pichai pushed back hard against what he called an “extraordinary” and “so far-reaching” demand by the U.S. government that could force Google to share its crown jewel: the data that powers its search engine.

Standing in a crisp dark suit on Wednesday, the 52-year-old tech executive warned that the Justice Department’s proposed remedies, aimed at dismantling Google’s alleged monopoly, would essentially amount to carving out the very core of the company.

“It would be a de facto divestiture,” Pichai said from the witness stand, contending that forcing Google to license its search data, even to competitors, would be akin to giving away decades of innovation and billions of dollars of investment at marginal cost.

That moment underscored just how high the stakes have become in the most consequential antitrust case the tech industry has seen in decades. Judge Amit Mehta has already ruled that Google illegally maintained its dominance in the search market. Now comes the hard part: deciding what to do about it.

The Justice Department, which filed its initial case in 2020, isn’t just looking for a slap on the wrist. It wants to unravel the mechanisms that have kept Google at the top for so long — exclusive contracts with phone makers like Apple and Samsung, the tight integration of Google’s search engine with its Chrome browser, and what it claims is a self-reinforcing loop where Google’s AI tools and search feed each other in a cycle that shuts out rivals.

Pichai, who has been at Google since 2004 and once oversaw the very browser the government now wants to be split off, told the court that dismantling such integration would deeply harm the company’s ability to fund future innovation. And nowhere, he said, is that innovation more important than in artificial intelligence.

“AI is one of the most profound technologies humans will ever work on,” he said, emphasizing Google’s $49 billion investment in the space. When he took over as CEO of Alphabet in 2019, Pichai said he pushed the company to become “AI first” — not just for search, but across all its products, reflecting a broader ambition: to evolve from delivering information to helping users take action.

However, the Justice Department argues that the very strategy has turned Google’s dominance in one area into a springboard for monopoly in another. Its lawyers say AI tools like Gemini benefit from Google’s search data, a trove no rival can match, and in turn funnel more users back into Google’s search ecosystem, creating what it called an “unbreakable feedback loop.”

That closed loop, the government contends, is especially dangerous now as generative AI takes center stage. OpenAI’s Nick Turley, who testified for the government earlier in the trial, said that after ChatGPT became a global phenomenon, his team realized they needed a search index to make their product more robust. But building one from scratch was too expensive. They approached Google, Turley testified, and got turned away.

That, in the Justice Department’s eyes, is the problem.

So they want a fix that cuts deeper than default settings and marketing agreements. They’re asking Judge Mehta to force Google to license its search data to competitors — a demand that Pichai portrayed as nothing less than a gutting of the company’s intellectual property.

However, government lawyers argue that without access to Google’s index, essentially a massive, constantly updated map of the internet — no AI model or rival search engine stands a real chance at competing. They point out that even when OpenAI sought workarounds, the imbalance in scale and resources remained insurmountable.

Still, Google’s defense hinges on the argument that competition exists and thrives, even in this landscape. Its lawyers argue that Meta’s Llama and OpenAI’s ChatGPT have leapfrogged Gemini in popularity. Pichai himself acknowledged as much on the stand, saying that while Google believes Gemini is a leading model, a “big gap” still exists between it and ChatGPT.

In court, the underlying message from Google has remained consistent: forcing the company to hand over the data that made it dominant would not only stifle innovation but also harm consumers who rely on its services every day.

The company’s critics — and there are many — see things differently. They say Google’s grip on search is precisely what has slowed innovation, blocked startups from flourishing and given the company unchecked power over how information is accessed globally.

But as the trial enters its final phase, with Google calling more executives and outside experts to its defense, including representatives from Mozilla and Apple, the big questions remain unresolved: Can a monopoly be undone without breaking the product itself? And what would the digital world look like if it were?

The court expects closing arguments by the end of May, and a ruling by August. Google, for its part, has indicated it will appeal — a sign that, whatever Judge Mehta rules, this legal showdown is far from over.

Tesla Board Denies Search for A New CEO Amid Poor Performance Buoyed by Elon Musk’s Politics

0

Tesla has dismissed reports that its board is actively seeking a replacement for CEO Elon Musk, calling the claims “absolutely false” even as the electric vehicle giant faces its sharpest decline in performance and investor confidence in years.

The company’s denial followed a Wall Street Journal report on Wednesday that claimed board members had approached executive search firms to begin scouting potential candidates to succeed Musk. The report, citing people familiar with the discussions, sent shares tumbling by as much as 3% in after-hours trading on platforms like Robinhood before regaining some ground.

Robyn Denholm, chair of Tesla’s board, pushed back strongly against the report, writing on social media platform X: “Earlier today, there was a media report erroneously claiming that the Tesla Board had contacted recruitment firms to initiate a CEO search at the company. This is absolutely false (and this was communicated to the media before the report was published). The CEO of Tesla is Elon Musk and the Board is highly confident in his ability to continue executing on the exciting growth plan ahead.”

Despite the denial, industry analysts say the report may have touched a nerve, reflecting mounting frustration from shareholders after a woeful start to 2025 that spilled over from last year. Tesla has seen its market value slashed by more than 30% since January, following dismal earnings, production slowdowns, and persistent concerns that Musk’s growing political involvement is undermining the company’s brand and alienating customers globally.

In the first quarter of 2025, Tesla posted a 9% drop in total revenue to $19.34 billion, far below the $21.11 billion expected by analysts, according to LSEG data. Automotive revenue fell 20% year-on-year to $14 billion, due to a combination of lower average selling prices, incentives, and temporary factory upgrades that interrupted Model Y production.

Net income collapsed 71% from the year before, plunging from $1.39 billion to just $409 million. Profit per share fell to 12 cents, down from 41 cents a year ago.

This steep decline has rattled investors, many of whom once believed Tesla could weather temporary dips in performance. Now, some point to Musk’s political entanglements, particularly his public alignment with Donald Trump, as a key factor dragging on the company’s value.

Musk’s political leanings have drawn increasing scrutiny. He was a cheerleader of the Trump campaign and was subsequently appointed to lead the Department of Government Efficiency (DOGE), which aims to streamline federal agencies’ spending and cut waste.

Although Musk said he would only spend “a day or two per week” on the project and later confirmed that he had stepped back from it, the damage may already have been done. His involvement in overt political activism, especially in a sharply divided global market, has risked alienating Tesla’s progressive consumer base, particularly in regions like California, Western Europe, and parts of Asia, where environmental and social concerns strongly influence EV adoption.

Analysts: Denial Is Not the End of the Story

Despite the board’s public defense of Musk, some analysts interpret the episode as a veiled message — a “warning shot” to the billionaire CEO that his position, while secure for now, is no longer untouchable.

Dan Ives of Wedbush Securities, a long-time Tesla bull, said the Wall Street Journal story, and the board’s quick rebuttal, reflect real tensions that have been brewing behind closed doors.

“On the WSJ/Musk article: While this was a very tense situation, we believe Musk clearly did the right thing [resigning from DOGE], and we believe Musk will remain CEO for at least five years at Tesla,” Ives said. “We would be surprised if the Board was still heading down this path. The Board did a warning shot.”

Others believe that while Musk’s recent actions have undoubtedly tested investor patience, there is still no clear successor with the vision, control, or charisma to replace him.

As Tesla struggles to regain its footing in a rapidly evolving EV industry, the leadership conversation is likely far from over. Competition is heating up from both legacy automakers and new entrants, and geopolitical tensions continue to complicate Tesla’s international growth strategy.

Analysts note that Musk’s ability to refocus on core operations while avoiding further political controversy will be critical in determining whether the company can rebound or continue to stumble.

Waymo and Toyota Team Up to Explore Self-Driving Tech for Personal Cars — Is Tesla Facing a New Threat?

0

Alphabet-owned Waymo and Japanese auto giant Toyota have entered into a preliminary partnership to explore a new frontier in autonomous driving: personally owned self-driving vehicles.

The alliance marks a potential strategic pivot away from the robotaxi-only playbook, hinting at a future where private consumers may soon own cars embedded with Waymo’s advanced driverless technology.

But beyond the immediate tech implications, the announcement has ignited debate on Wall Street and in the auto-tech sector, with some analysts now openly asking: Does Tesla finally have a real threat in the race for autonomous dominance?

Announced Tuesday, the collaboration between Waymo and Toyota aims to “leverage Waymo’s autonomous technology and Toyota’s vehicle expertise to enhance next-generation personally owned vehicles,” according to a joint statement. Though early in its formulation, the partnership hints at Toyota vehicles potentially being integrated into Waymo’s growing ride-hailing fleet, which has already been rolled out in Phoenix, San Francisco, Los Angeles, and most recently Austin.

Waymo’s co-CEO Tekedra Mawakana said the partnership could accelerate both companies’ ambitions in next-gen mobility. The idea isn’t just to fine-tune driver-assistance systems, but to potentially create fully autonomous vehicles that consumers can own — a concept once thought to be years away.

Analysts Divided: Is This Tesla’s Competition?

The announcement has not gone unnoticed on Wall Street, where opinions on its significance diverge sharply. Morgan Stanley, in a note to investors, said the Waymo–Toyota tie-up is a “major milestone” and represents “legit competition” for Tesla, which has long dominated the narrative around autonomous driving. The firm sees the partnership as a strategic alignment that brings together Google’s world-leading AI and mapping infrastructure with Toyota’s production scale and reliability.

But not everyone is buying into the buzz. Dan Ives, a tech analyst with Wedbush Securities and a long-time Tesla bull, downplayed the significance of the announcement.

“I disagree that the Waymo/Toyota is a groundbreaking deal and a threat to Tesla,” Ives said. “Tesla will own the autonomous market in my view, and no one can compete with their scale and scope. It starts in Austin in June, then the autonomous journey begins. Key chapter of growth.”

Tesla CEO Elon Musk has also been characteristically dismissive, suggesting in a recent earnings call that Waymo’s robotaxis are too expensive to be produced at scale, and reaffirming Tesla’s commitment to launching a fully autonomous ride-hailing service using its Model Y vehicles with the new “unsupervised” Full Self-Driving (FSD) software starting June in Austin.

Waymo Has the Lead — At Least for Now

However, by several measures, Waymo already has the upper hand when it comes to real-world deployment. The company is now delivering over 250,000 paid robotaxi rides per week, up from 200,000 in February. Its Waymo One service is active in four major metro areas — including Phoenix, San Francisco, Los Angeles, and Austin — with no human driver behind the wheel.

By contrast, Tesla’s so-called Full Self-Driving software, despite its name, still requires driver supervision and has yet to be validated as safe for fully autonomous operation. Regulatory agencies in the U.S. have investigated multiple incidents involving Tesla’s FSD and Autopilot features, further slowing Tesla’s path toward regulatory approval for a commercial driverless ride-hailing service.

Even Alphabet CEO Sundar Pichai acknowledged last week during Q1 earnings that Waymo has not finalized its long-term business model but emphasized the “optionality around personal ownership” as a promising revenue path. The Waymo–Toyota partnership appears to be a concrete step in that direction — potentially combining the tech know-how of Silicon Valley with the production might of Japan’s top automaker.

Not Waymo’s First Partnership — But Possibly the Most Significant

Waymo has previously collaborated with automakers including Jaguar Land Rover, Stellantis, Mercedes-Benz, Hyundai, and China’s Geely, often resulting in vehicles tailored for Waymo’s testing or ride-hailing fleets. However, those efforts largely remained limited in scale.

This Toyota deal may be different. Toyota is not only the world’s largest automaker by sales, but also a leader in hybrid and electrification platforms — a crucial advantage for developing energy-efficient autonomous systems. And unlike smaller OEMs, Toyota has the capacity to scale production globally if the partnership matures beyond testing.

Waymo said its collaboration with Toyota would not interfere with existing partnerships involving Hyundai and Geely for its Waymo One service. But industry insiders note that this tie-up could eventually open the door to mass-market personal AVs — a holy grail the industry has long chased but never achieved.

A Broader Shift in Industry Focus

Waymo and Toyota’s announcement also echoes a broader shift across the auto industry. Last year, General Motors paused its Cruise robotaxi operations after a series of safety concerns, refocusing instead on building self-driving systems for personal-use vehicles. Ford and Volkswagen similarly scaled back their investments in Argo AI, citing the difficulty of commercializing robotaxis at scale.

Meanwhile, Tesla remains an outlier, pursuing a vertically integrated approach and relying on camera-based systems, rather than the lidar and radar technologies used by Waymo and others. Musk has dismissed lidar as “unnecessary,” though Tesla’s software hasn’t proven itself in complex urban driving environments without human intervention.

It’s still unclear whether the Waymo–Toyota partnership will yield a commercially available autonomous personal vehicle. But if it does, it would mark a seismic shift — both in consumer mobility and in Tesla’s grip on the autonomous narrative.

Zenith Bank Posts N311.8 Billion Q1 Profit, Rides on Record Interest Income

0

Zenith Bank Plc, one of Nigeria’s largest lenders by assets, reported a record post-tax profit of N311.83 billion for the first quarter of 2025, a strong 20.7 percent year-on-year growth that cements its status as one of the country’s most profitable banks.

The performance was largely driven by a sharp rise in interest income, which hit N837.6 billion, the highest quarterly figure in the bank’s history.

The bank’s performance reflects a wider trend across Nigeria’s banking industry, where lenders are extending last year’s earnings momentum into 2025, thanks to sustained high interest rates, rising government borrowing, and improved asset quality.

Zenith’s interest income of N837.64 billion is a sharp 71.46 percent surge from the N488.7 billion posted in Q1 2024. This impressive growth in interest earnings, the money the bank makes from lending and investments in securities, powered the bank’s bottom line and accounted for nearly 90 percent of its gross earnings of N949.86 billion.

Zenith is not the only bank recording the massive growth. Other Tier-1 banks, including Access Holdings and GTCO, are also reporting double-digit profit growth in the first three months of 2025, driven largely by similar dynamics: strong loan expansion, bumper treasury bill investments, and rising customer deposits.

Zenith’s pre-tax profit rose 9.56 percent year-on-year to N350.82 billion, while its net interest income, which strips out the cost of funds, grew by a staggering 92.9 percent to N591.19 billion. This came despite a notable increase in interest expense, which climbed 35.34 percent to N246.45 billion, a reflection of the higher rates paid to attract and retain customer deposits.

Over 70 percent of interest expenses came from deposit liabilities, a consequence of the bank’s aggressive deposit mobilization in a high-rate environment. Customer deposits jumped by 35.14 percent year-on-year to N22.68 trillion, a boost of nearly N5.9 trillion in just three months.

However, even with higher funding costs, Zenith maintained robust margins. The bank’s net interest margin was supported by the efficient deployment of assets, with loans and advances to customers increasing by 16.19 percent to N10.05 trillion. Despite the rapid credit growth, impairment charges fell by 27.81 percent to N35.95 billion, pointing to an improvement in loan quality or stricter lending standards.

The balance sheet also swelled, with total assets climbing to N32.41 trillion, a 33.5 percent increase from Q1 2024. The expansion was largely driven by growth in deposits and increased investments in fixed-income securities, particularly Nigerian Treasury Bills, which Zenith ramped up by N2.68 trillion in the quarter. These instruments delivered a windfall — income from treasury bills alone jumped 113.24 percent to N328.8 billion, making up 39.23 percent of total interest income.

Zenith also earned N47.87 billion from cash balances placed with other banks, an uptick of 41 percent, though this remained the smallest of its three major interest income contributors.

In terms of non-interest income, Zenith posted only marginal gains. Overall, this segment rose just 0.98 percent year-on-year to N78.98 billion. Notably, electronic banking income dropped by 19 percent to N16.17 billion, likely due to changes in fee structures or reduced transaction volumes. But account maintenance charges rose 18.74 percent to N20.06 billion, partially offsetting the decline.

One notable outlier in the performance report is the drop in earnings per share (EPS), which fell by 7.66 percent to N7.59. The drop appears counterintuitive given the bank’s improved profitability, and is likely linked to the bank’s capital raise in late 2024 that increased the number of outstanding shares — a move aimed at bolstering its capital base amid planned expansion.

Also worth mentioning is the sharp rise in restricted deposits held with the Central Bank of Nigeria (CBN), which jumped by N2.24 trillion. This reflects the impact of the Cash Reserve Ratio (CRR) policy — a monetary tightening tool by the CBN that compels banks to leave a portion of their deposits with the regulator, limiting how much they can lend or invest.

Nonetheless, the bank retained strong liquidity. Cash and cash equivalents stood at N9.54 trillion, a 16.89 percent increase year-on-year.

Zenith’s Q1 performance continues a trend first established in 2023 when Nigerian banks began reporting windfall profits as interest rates soared and Treasury yields became more attractive. The Central Bank’s tighter monetary stance aimed at taming inflation and supporting the naira has worked to the advantage of lenders, even as other sectors grapple with high borrowing costs.

Analysts expect Zenith and its peers to continue riding this wave in the short to medium term, especially if the policy environment remains restrictive. However, they expect the longer-term outlook to depend on how banks manage credit risk as they expand their loan books amid a fragile economy.