Bloomberg Tech
2026-04-28 · Hosted by Caroline Hyde, Ed Ludlow · Bloomberg / iHeartMedia
Executive Summary
The dominant story of the episode is OpenAI’s agreement to end its exclusivity arrangement with Microsoft, freeing OpenAI to distribute and sell its models across all major cloud platforms including AWS, while Microsoft retains free access to the models through 2032 and drops its revenue-share obligation to OpenAI. The initial market reaction was dramatic — Microsoft fell 4% in pre-market before recovering to down roughly 0.4%, and Amazon spiked on the assumption it would be a primary beneficiary before giving back those gains. Three additional AI-centric stories drove the session: China’s NDRC unilaterally blocked Meta’s completed $2 billion acquisition of Manus AI on national-security grounds, a closely watched analyst report claimed OpenAI is working with Qualcomm on a future smartphone device (sending Qualcomm up 14% in pre-market before closing down roughly 1%), and the Elon Musk vs. OpenAI trial — seeking up to $134 billion in damages — began jury selection in Oakland. Against this backdrop, Meta and Microsoft are each heading into Wednesday earnings reporting tens of thousands of combined job cuts while simultaneously projecting record AI capital expenditure, a tension analysts framed as the defining paradox of the current tech cycle. The Philadelphia Semiconductor Index, which had posted an 18-session record win streak, gave back roughly 2% on the day.
Key Stories & Changes
1. OpenAI Ends Azure Exclusivity — Microsoft and Amazon React
OpenAI agreed to terminate the exclusivity clause that had kept its models available solely on Microsoft Azure
Microsoft retains free access to OpenAI models through 2032; in exchange, Microsoft is no longer paying OpenAI a revenue share on revenues derived from OpenAI technology
OpenAI is now free to distribute and sell models to all major cloud providers; an AWS partnership had already been struck in recent months (reported as January–February), which Microsoft was reportedly unhappy about
Microsoft stock fell 4% in pre-market on the news, recovered to approximately -0.4% by midday
Amazon shares spiked in pre-market on the assumption it would benefit, then largely gave back the move
Bloomberg Intelligence analyst Anna Ragnar described the amended partnership as a net positive for enterprise AI adoption, and specifically for AWS
Brody Ford (Bloomberg reporter) noted the financial complexity: OpenAI is simultaneously a customer of Microsoft, pays Microsoft for services, and is Microsoft’s largest investment holding
2. China Blocks Meta’s $2 Billion Manus AI Acquisition
China’s National Development and Reform Commission (NDRC) issued a one-sentence order blocking Meta’s $2 billion acquisition of AI startup Manus
The deal had been announced in December 2025; Manus’s founders were originally based in mainland China before moving to Singapore mid-2024 and declaring it a Singapore company
The NDRC acted without naming Meta directly and provided no detailed legal rationale
Key complication: Meta employees had already integrated with Manus, executives were cooperating, and all Manus investors had already received payment — making it unclear how a true unwind is enforceable
Framing from Peter Elström (Bloomberg Asia Tech editor): Beijing has been pressing Chinese AI companies not to accept US investment; the move reflects concern over “losing critical AI technology” to the US
Political context: President Trump is expected to meet Xi Jinping in coming weeks; semiconductor chip access for China remains the primary negotiating chip, though software/AI may be on the table
Meta stated the deal was done in compliance with applicable laws and said it expects a resolution; stock was up ~6.1% on the session
Jacob Hellberg (US Undersecretary of State for Economic Affairs) interpreted the move as an example of China’s “economic coercion” and said the US is addressing supply-chain concentration risk through the Paxillica coalition (14 countries) and a new “forward deployed industrial base” partnership with the Philippines announced the prior week
3. Qualcomm Surges and Reverses on OpenAI Smartphone Analyst Report
Ming-Chi Kuo of TF International Securities posted a report on X claiming Qualcomm (and MediaTek) is working with OpenAI as the silicon provider for a future smartphone device
Qualcomm jumped 14% in pre-market, opened up 8%, and then reversed to down roughly 1% by midday
Mass production reportedly not expected until 2028 — making it extremely difficult to price into current valuation
None of the companies involved confirmed or commented to Bloomberg
Context from Ryan Vistellaker (Bloomberg equity reporter): Qualcomm faces near-term headwinds from rising memory prices weighing on handset chip demand, and Apple is developing its own modem in-house — removing a major customer
Tim Arcuri (UBS Semi/SemiCap analyst, Global Co-Head of AI): noted Qualcomm has a credible server/agentic effort; cautioned that “OpenAI tends to do a lot of deals with a lot of different companies” — warranting a discount on any single announcement; flagged that 70% of Qualcomm’s operating profit comes from phones, and Apple may stop paying licensing fees once its modem is fully in-house; Qualcomm earnings expected Wednesday
4. Musk vs. OpenAI Trial — Jury Selection Begins
Jury selection began in Oakland for Elon Musk’s lawsuit against OpenAI, co-founders Sam Altman and Greg Brockman, and Microsoft
Musk is seeking up to $134 billion in damages, which he says should go to OpenAI’s charitable arm
Only two claims remain (others dropped or dismissed): (1) Altman and OpenAI violated a promise to maintain a permanent charitable mission to develop safe, open-source AI for public benefit — violated when OpenAI created a for-profit affiliate in 2019; (2) Altman and OpenAI received undeserved benefits (including Musk’s investment) because of these broken promises
Remedies sought include pushing Altman out as CEO and unwinding the for-profit conversion completed early 2025
Key witnesses scheduled to testify: Sam Altman, Elon Musk, Mira Murati (former OpenAI CTO, now Thinking Machines), Satya Nadella (Microsoft CEO)
Dorothy Lund (Columbia Law School professor): the jury is advisory only — Judge Gonzalez Rogers is not bound to follow the verdict; noted Musk’s conflict of interest as founder of rival xAI (now owned by SpaceX), which would benefit if the suit succeeds (as disclosed in SpaceX’s IPO letter); called the case “easily viewed cynically” and said the judge has already raised Musk’s motives in pre-trial motions
Lund assessed Musk’s argument that the case would set deterrent precedent for other nonprofits as “weak” — corporate law is enabling by design and allows businesses to change form based on business needs
5. Meta and Microsoft Earnings Preview — Layoffs vs. AI Capex Paradox
Wednesday earnings: Alphabet, Microsoft, Amazon, and Meta all report after the bell — combined market cap context framed as a $16 trillion test for the market rally
Both Meta and Microsoft signaled cuts days before earnings:
Meta: planning to slash approximately 10% of its workforce
Microsoft: offering buyouts “on a scale it has never attempted before”
Combined potential: up to 23,000 roles
Simultaneously, the group is projected to spend more than $640 billion on capital expenditures in 2026, up from more than $400 billion the prior year
Mag 7 earnings growth projected at 19%
Amazon free cash flow expected to be negative this quarter — potentially the widest since 2022 (when it was investing in pandemic-era warehouse expansion)
Meta Q1 free cash flow expected to be the smallest in nearly four years
Investors are focused on cloud sales (AWS, Google Cloud) as the primary proof point for AI ROI
Sarah Franklin (Lattice CEO): warned that cutting to free up capex is a “short-term gain on the spreadsheet, long-term loss in workforce knowledge”; noted “token maxing” emerging as a productivity metric but cautioned it is not a reliable performance proxy; distinguished Meta’s hard cuts from Microsoft’s voluntary buyout model
6. Semiconductor Sector — Record 18-Day SOX Streak Gives Back Ground
The Philadelphia Semiconductor Index (SOX) had ended the prior Friday up for a record 18th consecutive session; the index pulled back approximately 2% on April 27
Tim Arcuri (UBS) attributed the streak partly to a resurgence of the analog semiconductor sector: Texas Instruments (TI) gave strong guidance and bullish commentary on capacity constraints in the industrial segment, prompting customers to begin building inventory buffers
Arcuri also cited bullishness on semiconductor equipment (KLA mentioned), with expectations of further guidance raises later in the week; noted “so much capacity has to get built out over the next three years”
Intel went to the investment-grade bond market on April 27 to help finance its Ireland buildout; Arcuri said Intel’s foundry turnaround is progressing faster than its product roadmap — product side remains the key caution; more foundry customer announcements expected in fall 2026
Bitcoin trading at $76,600 per token during the episode; the hosts described the overall session as a “risk-off environment”
7. Tech Briefs — Space Solar, TSMC IP Theft, and AI Startup Funding
Meta reserved up to 1 gigawatt of power from startup Overview Energy, which plans to beam solar energy from orbit back to Earth; Meta aims to secure uninterrupted data center energy supply by 2030
A former Tokyo Electron engineer was sentenced to 10 years in prison in Taiwan for stealing TSMC’s proprietary data; Taiwan cited the case as part of heightened efforts to guard its semiconductor IP
Former DeepMind researcher and AlphaGo architect David Silver raised $1.1 billion for new startup Ineffable Intelligence at a $5.1 billion valuation; backers include Sequoia, Nvidia, and Google; focus is on reinforcement learning and robotics rather than large language models
Trends Identified
1. OpenAI’s Expanding Multi-Cloud Ambition
The termination of OpenAI’s Azure exclusivity marks the latest step in OpenAI’s transformation from a Microsoft-dependent partner into an independent platform company seeking to reach customers through every major cloud. The AWS deal earlier this year was reportedly the catalyst for Microsoft renegotiating the original exclusivity terms, and the price Microsoft extracted — no more revenue share — illustrates the shift in leverage. For OpenAI, constrained compute and the need to reach as large an addressable market as possible make multi-cloud distribution a strategic necessity, even if it means ceding exclusivity to its most important investor.
2. China Using Regulatory Leverage Over AI Technology Transfer
Beijing’s move to block the Meta-Manus deal — a completed transaction involving a company that had formally relocated to Singapore — signals a new and more aggressive posture on AI technology export control. Chinese regulators are pressing domestic AI companies not to take US capital, and the NDRC intervention suggests Beijing is willing to act retroactively even when deals appear legally structured to avoid Chinese jurisdiction. With Trump-Xi negotiations imminent, the Manus block appears designed to maximize China’s leverage on AI and chip access simultaneously.
3. AI Hardware Speculation Cycle — Hype, Reversal, and Earnings Reality
The Qualcomm episode is a textbook illustration of how AI hardware speculation can drive dramatic intraday swings that quickly reverse when valuation reality sets in. A single analyst post on X moved Qualcomm up 14% in pre-market before the stock closed in the red. The pattern reflects genuine investor anxiety about who will supply the physical layer of AI devices — and equally genuine uncertainty about timing, revenue magnitude, and competitive dynamics. With Qualcomm already under structural pressure from Apple’s in-house modem and memory-price headwinds, the market is pricing optionality on AI hardware partnerships without the data to do so rigorously.
4. The Tech Industry’s AI Restructuring Trade-Off
Meta and Microsoft’s simultaneous layoff announcements and record capex pledges crystallize a structural shift under way across the industry: human capital is being traded for compute capital. The distinction between cutting headcount to fund AI investment versus the pandemic-era cuts to correct over-hiring is real but narrow in practice — both involve experienced people leaving organizations. What is new in the current cycle is that the capex commitments ($640 billion projected for the Mag 7 group in 2026) are so large relative to current revenue that free cash flow is expected to turn negative for Amazon and hit multi-year lows for Meta, making Wednesday’s earnings a genuine test of whether investors will continue to fund the trade-off thesis.
5. US-China Tech Decoupling as Active Statecraft
The Hellberg interview framed what is often described as supply-chain resilience in explicitly competitive terms: the US is building Paxillica, a 14-country economic security coalition, and deploying a “forward industrial base” with the Philippines, explicitly designed to counter China’s Belt and Road connectivity model with private-sector-led manufacturing partnerships. The emphasis on de-risking “single points of failure” — helium for chip fab, rare-earth magnets, logistics nodes — reflects how the Iran conflict exposed practical vulnerabilities in the AI supply chain that had previously been theoretical.
6. Musk-Altman Trial as a Test of AI Governance Norms
The case going to trial moves what had been a social-media feud between competing AI founders into a venue that could produce legal precedent on the obligations of AI companies that began as nonprofits. Professor Lund assessed the precedent risk as limited — corporate law is enabling, not restrictive, and OpenAI’s governance history is unique — but the case will nevertheless put on public record the internal decision-making behind OpenAI’s for-profit conversion, the nature of Musk’s early contributions, and the extent to which the judge views the lawsuit as a competitive weapon wielded by a rival AI company founder. —-
Sentiment Analysis
Overall Market Sentiment: Cautious / Risk-Off
An extraordinary volume of AI-related news headlines hit simultaneously, producing dramatic intraday swings across Microsoft, Amazon, Qualcomm, and Meta that mostly reversed by midday. The SOX gave back 2% of its record streak and Bitcoin traded at $76,600, consistent with a modest risk-off tone.
Risk Factors Highlighted
OpenAI multi-cloud transition execution risk: As OpenAI distributes models across AWS, Azure, and potentially other platforms, Microsoft’s historical integration advantage erodes — but OpenAI faces new complexity managing multiple hyperscaler relationships simultaneously.
China’s expanding extraterritorial AI enforcement: The Manus block demonstrates Beijing’s willingness to act on deals involving Chinese-founded companies even after formal relocation to third-party jurisdictions, creating uncertainty for any cross-border AI M&A with Chinese-origin founding teams.
Qualcomm structural dependency on handsets: With 70% of operating profit from phones, rising memory prices, and Apple developing its own modem in-house, Qualcomm’s near-term earnings trajectory is under pressure regardless of long-term AI hardware optionality.
Mag 7 capex / free cash flow tension: Projected spend of over $640 billion in capex in 2026 (up from $400 billion) is expected to produce negative free cash flow at Amazon and multi-year lows at Meta — raising the question of whether return on AI investment will materialize fast enough to satisfy investors.
OpenAI non-profit conversion legal exposure: If Judge Gonzalez Rogers finds against OpenAI, remedies could include unwinding the for-profit conversion completed in early 2025 and removing Altman as CEO — creating existential governance uncertainty at the most influential AI lab in the world.
Musk conflict-of-interest creating litigation distraction for xAI and OpenAI: Both companies face reputational risk and management distraction from a public trial; xAI’s IPO documents disclose a financial interest in the lawsuit’s outcome, which could color public and regulatory perceptions of its motives.
AI workforce transition cultural cost: Sarah Franklin (Lattice) explicitly flagged that simultaneous layoffs at Meta and Microsoft will damage culture and institutional knowledge even if the spreadsheet math favors cutting; the risk is a “long-term loss in w
This episode was covered in today’s The Market Signal — 2026-04-28, a cross-source synthesis of multiple podcast reports.