DD
MM
YYYY

PAGES

DD
MM
YYYY

spot_img

PAGES

Home Blog Page 10

Tekedia Capital Invests in OpenSec, AI Spec Framework Startup

0

Tekedia Capital is excited to announce our investment in OpenSpec, one of the leading open-source specification frameworks for AI agents. OpenSpec helps teams capture what they are building, why they are building it, and how it should work before a coding agent writes a single line of code. In a rapidly growing ecosystem of AI-native software development, OpenSpec is emerging as an important layer of infrastructure.

Why did we invest? Two ideas influenced our thinking.

The first comes from the legendary Chief Dr. Sir Oliver De Coque, who sang, “Egwu ?ma si na Chi” (good music comes from God). Beyond the poetry of the lyrics lies a deeper truth: before beautiful music is performed, someone must first organize the notes. There must be a structure, a design, and a framework. Great art does not emerge from randomness; it emerges from thoughtful architecture. Sir Oliver posited that his God was his framework.

The second comes from engineering. In my experience as a mixed signal design engineer, the most difficult part of building a product is defining the specification. The challenge is not necessarily connecting transistors or writing code; it is clearly describing what the final system must do, to enable what the marketing team is expecting!

Why is that important? Imagine ten engineering teams distributed across different locations working on the same product. The power team, signal-processing team, analog team, verification team, and software team cannot wait for one another to finish before starting work. They must operate in parallel where necessary. The only way that works is when everyone is guided by a common specification. The specification becomes the contract that ensures that when all the pieces come together, a coherent product emerges!

The AI era has the same requirement. Today, anyone can prompt. Anyone can generate code. Anyone can launch an agent. But the real question is this: how do you ensure that the output matches the intention? How do you move from prompting to engineering?

OpenSpec addresses that challenge by introducing discipline before generation. It enables teams to do the hard work of defining requirements, workflows, objectives, constraints, and expected outcomes before the AI begins execution. Instead of generating code first and spending weeks correcting mistakes, teams can establish clarity upfront and dramatically improve productivity. We like this mission, the product (tens of thousands of users), and the future it enables. Hence the cheque!

Roche Chairman Severin Schwan Blasts U.S. Tariff Tactics as “Cold-Blooded Blackmail,”

0

Severin Schwan, chairman of Swiss pharmaceutical powerhouse Roche, delivered a rare and sharply critical assessment of U.S. trade policy on Thursday, describing Washington’s use of tariff threats to force drug price reductions as “cold-blooded blackmail.”

His comments shed light on the intense pressure major international drugmakers face amid the Trump administration’s aggressive push to reshape the pharmaceutical trade industry. Schwan was referring to a late-2025 agreement in which Roche agreed to significantly lower prices for its medicines in the United States after U.S. officials threatened to impose steep tariffs, reportedly as high as 200%, on pharmaceutical imports.

Speaking at an event in the Swiss city of Interlaken, he made no effort to soften his view of the negotiations.

“If someone points a gun at you and says ‘if you don’t sign, there’ll be 200% tariffs tomorrow’, I wouldn’t necessarily describe that as a deal. So in a legal sense that’s perhaps an agreement, but it’s basically cold-blooded blackmail,” he said.

The remarks expose the growing friction between global pharmaceutical companies and the Trump administration, which has repeatedly argued that foreign firms have taken advantage of the U.S. market to charge American consumers excessively high prices. By wielding the threat of tariffs, the administration has sought both immediate price concessions and longer-term shifts toward increased domestic drug manufacturing.

Beyond the immediate dispute, Schwan identified rising protectionism from both the United States and China as Roche’s single biggest geopolitical concern. This assessment carries significant weight given Roche’s global footprint. As one of the world’s largest biotech and diagnostics companies, Roche operates extensive research, development, and manufacturing networks across Europe, North America, and Asia.

Fragmentation of global trade rules, export controls on critical materials, and retaliatory measures threaten to disrupt these carefully optimized supply chains and slow the pace of innovation.

The U.S.-China strategic rivalry has already complicated matters for the industry. Beijing has tightened controls on rare earth minerals and pharmaceutical ingredients, while Washington has expanded scrutiny of supply chain security and intellectual property practices.

For a company like Roche, which relies on cross-border collaboration for everything from clinical trials to advanced manufacturing, this dual pressure from the world’s two largest economies creates a challenging operating environment.

Schwan’s unusually candid criticism reflects broader frustration across the European and international pharmaceutical sector. Many companies have quietly accepted price cuts or localized production commitments to avoid tariffs, but executives worry about the long-term consequences.

Forced price reductions in the lucrative U.S. market, which often accounts for a disproportionate share of global profits for innovative medicines, could reduce the financial resources available for research and development. This is particularly concerning at a time when the industry is investing heavily in next-generation therapies such as antibody-drug conjugates, personalized medicine, and advanced diagnostics.

The Trump administration’s approach, while politically popular domestically, risks undermining the incentive structure that has driven decades of medical progress. Pharmaceutical innovation is extraordinarily expensive and risky, with the vast majority of development programs failing to reach the market. High prices in the United States have historically helped subsidize global R&D.

Industry leaders have expressed concern that significant and sustained price erosion could slow the pipeline of new treatments for cancer, Alzheimer’s, rare diseases, and other conditions.

Roche has long positioned itself as a leader in oncology, immunology, and diagnostics. The company’s ability to maintain robust innovation depends on stable and predictable access to major markets. Thus, Schwan is suggesting that while Roche complied with U.S. demands to avoid tariffs, the experience has left a sour taste and heightened concerns about future unpredictability in trade policy.

The episode also illustrates the limited leverage many foreign companies have when dealing with the U.S. market. With America representing roughly 40-50% of global pharmaceutical revenue for many innovative drugs, walking away from negotiations is rarely a realistic option. This dynamic gives Washington considerable negotiating power but also risks straining transatlantic relations and encouraging companies to accelerate supply chain diversification — moves that carry their own costs.

Looking ahead, pharmaceutical companies are likely to adopt a multi-pronged strategy: engaging constructively with U.S. policymakers where possible, while quietly accelerating efforts to localize production, diversify supply chains, and strengthen relationships in other growth markets such as Europe, China, and emerging economies.

For Roche specifically, the focus remains on executing its rich pipeline and maintaining leadership in key therapeutic areas. However, Schwan’s remarks serve as a public signal that the company, and by extension much of the industry, views the current trade environment as unsustainable and potentially damaging to long-term innovation.

 

Poke Becomes First AI Agent Approved for Apple’s Messages for Business Platform

0

Poke, a startup that simplifies interaction with powerful AI agents to the simplicity of sending a text message, has secured approval to operate on Apple’s Messages for Business platform.

This marks the first time a standalone third-party AI agent has been integrated into the service, previously reserved for businesses such as airlines, retailers, and hotel chains to communicate directly with customers via iMessage.

Launched in March, Poke aims to make advanced AI accessible to everyday users who lack technical expertise or interest in complex interfaces. Through simple text conversations, users can manage daily planning, calendars, health and fitness tracking, smart home controls, and photo editing. To date, the service has relayed more than 100 million messages across SMS, Telegram, WhatsApp (in select markets), and now iMessage.

The integration expands Poke’s reach within Apple’s tightly controlled ecosystem and comes just days before Apple’s Worldwide Developers Conference (WWDC), where the company is expected to unveil an AI-optimized Siri and other developer tools. While rumors suggest Apple may open its App Store to AI agents, Poke’s approval operates through the existing Messages for Business framework, allowing users to interact with the AI directly through iMessage without downloading a separate app.

Marvin von Hagen, co-founder of The Interaction Company (the Palo Alto-based startup behind Poke), highlighted the strategic importance of the move. The company will pay Apple on a per-user basis — a structure von Hagen described as significantly more favorable than recent fee increases imposed by Meta on its WhatsApp platform due to EU regulations.

This revenue-sharing model could represent a meaningful new income stream for Apple while giving AI startups like Poke scalable distribution within one of the world’s most valuable consumer ecosystems.

“I think that Apple is just noticing this is the best way to offer AI, and … actually, good for them, because they charge us. They charge us per user on the platform and actually make money with this, especially if it becomes really big,” von Hagen said.

Approval Process and Trust Factors

Securing Apple’s approval was no small feat. It required Poke to demonstrate robust live human support capabilities as a fallback, clearly identify itself as an AI agent, and adhere to strict design guidelines. This included using Apple-style buttons, interface elements, and link previews instead of inline links.

Von Hagen noted the process took several months and will likely pose a similar barrier for other AI agents seeking entry.

“This took a couple of months to adhere to all of these standards, and it will take anyone else who wants to build on this — it will also take them a couple of months to get through this approval process,” he said.

Trust played a significant role in Poke being first. Von Hagen emphasized that the startup’s focus on quality and long-term brand integrity, rather than aggressive growth tactics, aligned well with Apple’s standards.

Poke is currently rolling out invites to existing users, giving them the option to shift interactions to iMessage. The service will initially remain free for businesses and consumers, with paid tiers expected soon.

This development is significant on multiple levels. For Apple, it represents a pragmatic step toward embracing AI agents without immediately opening the App Store or core iOS systems. By integrating capable third-party agents into Messages for Business, Apple can enhance the utility of its messaging platform while generating new revenue and gathering valuable data on AI usage patterns.

For the broader AI industry, Poke’s approval highlights the growing push toward agentic AI — systems that go beyond answering questions to actively performing tasks. While giants like OpenAI, Anthropic, and Google dominate headlines with frontier models, startups like Poke are focusing on usability and accessibility. Poke is attempting to meet users where they already are, rather than forcing them into new apps or interfaces by operating across SMS, messaging apps, and now iMessage.

With WWDC approaching, Apple is under pressure to demonstrate meaningful AI progress. Allowing Poke onto Messages for Business could serve as an early signal of how Apple envisions third-party AI integration — controlled, secure, and revenue-generating for the company.

Poke, backed by Spark Capital, General Catalyst, and several angels, recently raised an additional $10 million, bringing its total funding to $25 million. The 10-person startup is now valued at $300 million post-money. Its lean structure and focus on practical, text-first interactions differentiate it from more resource-intensive agent platforms.

Von Hagen and his team are betting that simplicity and seamless integration will drive adoption. In a world where many AI tools still feel experimental or overly complex, Poke’s text-message interface lowers the barrier dramatically. Early traction, 100 million messages in just a few months, suggests strong product-market fit among users seeking frictionless AI assistance.

Foxconn and Intel Forge AI Infrastructure Alliance, to Build Next-Gen AI Systems

0

Taiwanese manufacturing giant Foxconn is deepening its push into artificial intelligence infrastructure through a new partnership with Intel, a move that highlights how the AI boom is reshaping the competitive landscape of the semiconductor and data center industries.

Foxconn, formally known as Hon Hai Precision Industry, announced Thursday that it will collaborate with Intel to jointly develop and deploy next-generation AI infrastructure and intelligent computing platforms.

The partnership comes as technology companies, cloud providers, and governments worldwide pour unprecedented sums into AI computing capacity. Industry executives estimate that hyperscalers and major cloud providers will spend close to $1 trillion annually on AI infrastructure within the next few years, creating enormous opportunities for chipmakers, server manufacturers, and data center equipment suppliers.

Under the agreement, Intel will contribute its processor and accelerator technologies, while Foxconn will leverage its vast manufacturing footprint and system-integration capabilities. The companies plan to collaborate on AI data center equipment, including server racks powered by Intel’s Xeon processors and AI accelerator chips.

The partnership will also focus on several increasingly important areas of AI infrastructure, including high-speed interconnect technologies, advanced cooling systems, and energy-efficiency solutions.

Those areas have become critical battlegrounds in the AI race. As AI models grow larger and more computationally demanding, the challenge is no longer simply producing faster chips. Companies now must also solve problems involving power consumption, heat dissipation, and data movement between thousands of interconnected processors.

This is where Foxconn sees a major opportunity.

Traditionally known as the world’s largest contract electronics manufacturer and the primary assembler of products for companies such as Apple, Foxconn has been rapidly repositioning itself as a supplier of AI infrastructure. The company is already the largest manufacturer of AI servers for Nvidia and has become one of the biggest beneficiaries of the global AI investment wave.

Chairman and Chief Executive Officer Young Liu recently said that spending by major cloud providers represents one of the strongest growth drivers in the company’s history.

“Our collaboration with Intel will combine the strengths of both companies across computing platforms, system integration, and global supply chain capabilities,” Liu said in a statement.

For Intel, the partnership represents another effort to strengthen its position in the rapidly evolving AI ecosystem. Although Nvidia remains the dominant force in AI accelerators, Intel has been expanding its presence in AI infrastructure through Xeon processors, accelerator technologies, and advanced packaging capabilities. The company has also benefited from a growing shortage of high-performance CPUs, which remain essential for AI workloads alongside graphics processors.

Notably, the partnership extends beyond traditional data centers. Foxconn and Intel said they plan to develop AI systems for factories, smart cities, and robotics applications, reflecting a broader industry shift toward “edge AI,” where intelligence is deployed directly into devices and industrial environments rather than solely through centralized cloud infrastructure.

That opportunity could prove enormous.

As enterprises increasingly adopt autonomous systems, industrial robots, and AI-powered automation, demand is expected to grow for compact computing systems capable of running sophisticated AI models outside conventional server farms.

The companies also disclosed plans to explore custom chip development and broader system integration solutions, suggesting the alliance could eventually move into the rapidly expanding market for bespoke AI semiconductors.

Custom silicon has become one of the hottest segments of the semiconductor industry, with companies such as Alphabet, Amazon, and Microsoft increasingly designing their own processors to optimize performance and reduce costs.

The Foxconn-Intel collaboration arrives as competition intensifies across every layer of the AI infrastructure stack. Companies are racing not only to build more powerful chips but also to secure positions in server manufacturing, networking equipment, cooling technologies, power systems, and AI deployment platforms.

Neither company disclosed the financial value of the agreement, identified customers, or provided a timeline for commercial deployment. Nevertheless, the announcement shows that AI infrastructure is evolving into a massive ecosystem that requires integrating chips, servers, networking, software, and manufacturing expertise.

The deal thus reinforces Foxconn’s transformation from a contract manufacturer into a strategic AI infrastructure player, while it provides another avenue to expand Intel’s footprint in a market being defined by demand for large-scale AI computing systems.

Analysts expect partnerships such as this to become common as companies seek to combine technological expertise with manufacturing scale to capture a share of one of the industry, especially as spending on AI infrastructure accelerates globally.

Meta Launches Business AI Agent with Agentic Capabilities, Signaling Deeper Push into Enterprise Market

0

Meta Platforms on Wednesday introduced a new artificial intelligence agent designed to handle day-to-day business operations, marking a significant step in the social media giant’s effort to expand beyond consumer apps into enterprise AI tools.

Unveiled at the company’s WhatsApp-focused Conversations conference in London, the Business Agent adds “agentic” functionality — the ability to autonomously take actions on behalf of businesses, such as booking calendar appointments, qualifying leads, closing sales, and processing payments.

This goes well beyond simple rule-based chatbots by enabling more complex, multi-step workflows.

Naomi Gleit, Meta’s head of product for the initiative, described the launch as a clear enterprise play.

“This is definitely an enterprise play,” he said.

Meta said more than 1 million businesses are already using earlier chatbot versions of such agents on WhatsApp and Messenger. The new version will be rolled out globally to businesses of all sizes and added to Instagram as well. Initially free, paid subscription options are planned for the coming months.

Broader Business Agent Platform and Ecosystem Integration

Alongside the in-app Business Agent, Meta is launching a wider Business Agent Platform that allows companies to build and deploy custom AI agents across their operations. The platform connects to hundreds of non-Meta systems, including Shopify, Zendesk, and Shopee, and offers enterprise-grade controls, guardrails, and performance measurement tools.

This platform approach aims to position Meta as a central orchestrator in the growing agentic AI space, where AI systems can independently manage workflows rather than just respond to queries.

Gleit highlighted the need for a unified experience, saying: “The number one thing I hear, especially from small businesses, is ‘I just want to go to one place that can do all the things.’”

She is leading a new Enterprise Solutions team as part of a recent company-wide restructuring around AI. The team will deploy squads of forward-deployed engineers to work directly with large customers, a model popularized by companies like Anthropic, to navigate internal adoption challenges and customize solutions.

Meta is also working to consolidate its various AI agents, including internal workflow tools, the public Meta AI bot, and a recently launched ads-focused business assistant.

The launch spins off Meta’s ambition to leverage its massive reach across WhatsApp, Messenger, and Instagram, platforms with billions of users, to gain ground in the enterprise AI market against specialists like OpenAI, Anthropic, and Google.

By embedding agentic capabilities directly into the communication tools businesses already use to engage customers, Meta is betting it can become a one-stop platform for both consumer-facing interactions and internal operations. This strategy could help the company diversify revenue beyond advertising while deepening its integration into business workflows.

Gleit emphasized the importance of orchestration and efficiency.

“We actually want to take actions now. We actually want it to be able to complete the payment, to process the booking, to place the order,” she said.

The move comes as competition in agentic AI intensifies. Rivals are rapidly advancing their own autonomous agents, while Microsoft and Apple are enhancing their ecosystems with similar capabilities.

However, Gleit acknowledged the risks of permitting AI agents to act on behalf of businesses, particularly around security and reliability. Meta recently faced an embarrassing incident in which hackers tricked its AI support chatbot into granting access to high-profile Instagram accounts.

She noted the issue stemmed from a flawed technical check rather than the agent itself.

“It wasn’t the agent. The agent actually exposed a technical check that wasn’t working. There was sort of a separate system and technical check that had a bug, and because people were using the agent, they discovered it,” she said.

Shares of Meta rose more than 3% in morning trading, suggesting investors view the announcement positively as a step toward new revenue streams and deeper platform stickiness.

The Business Agent and Platform represent Meta’s latest attempt to monetize its vast messaging ecosystem while positioning itself in the high-growth enterprise AI sector.

As AI agents move from experimental tools to core business infrastructure, Meta’s entry adds significant competition and choice for companies looking to automate operations. But the move represents more than just a product launch for Meta — it is seen as part of a broader strategic shift to build durable, high-margin AI businesses that complement its advertising empire and reduce reliance on any single revenue source.