AI News: May 28, 2026 — ByteDance Doubles Down on AI Infrastructure, OpenAI's IPO Is Now Official

1. ByteDance Is Weighing Up to $70 Billion in AI Infrastructure Spending This Year

Bloomberg reported on May 27 that ByteDance — the parent company of TikTok and Douyin — is discussing capital expenditures of as much as $70 billion in 2026 as it expands data centers and AI infrastructure. The company would fund most of that spending from the roughly $50 billion of profit it earned in 2025. The figure, which Bloomberg notes is preliminary and subject to quarterly adjustment, would represent more than double ByteDance's prior-year spending and put it alongside the largest US hyperscalers in absolute terms.

The spending rationale is straightforward. ByteDance's AI assistant Doubao has been growing rapidly in China, and the company has explicit ambitions to challenge US AI leaders internationally. It recently struck a deal to acquire millions of Qualcomm chips to support its agentic AI services. Crucially, data center construction costs in China are significantly lower than in the US, meaning ByteDance may be able to build equivalent compute capacity with lower nominal spending than a Meta or Microsoft operating domestically.

For context: Amazon is projecting approximately $200 billion in capex in 2026, Alphabet is targeting $175–185 billion, Meta is at $115–135 billion, and Microsoft is tracking toward a similar range. ByteDance at $70 billion would sit just below this US hyperscaler tier. Tencent's 2025 capex was 79.2 billion yuan (~$11 billion); Alibaba's was 126 billion yuan (~$17 billion) for its most recent fiscal year. The Chinese AI infrastructure buildout is accelerating sharply across all three companies, with ByteDance now the most aggressive spender of the three.

2. OpenAI Confidentially Filed Its IPO S-1 with the SEC on May 22

Fortune, CNBC, Reuters, Axios, and Bloomberg all confirmed on May 22 that OpenAI filed a confidential draft S-1 registration statement with the US Securities and Exchange Commission, formally kicking off the IPO process. Goldman Sachs and Morgan Stanley are leading the deal. The company is targeting a public market debut in Q4 2026, with September as the early target date. Analysts covering the deal expect the listing to push OpenAI past a $1 trillion market capitalization, above its current private-market mark of $852 billion.

A confidential filing is the quiet first move in a months-long process. The S-1 stays sealed until approximately 15 days before the public roadshow — meaning detailed financials, including the Q1 2026 revenue and loss figures, will not be publicly visible until shortly before pricing day. The SEC typically provides feedback within 30 days of a confidential filing. After that comes the unredacted S-1 release, the roadshow, and finally pricing and listing.

The filing arrives despite a difficult profitability picture: OpenAI is reported to be losing $1.22 for every $1 of revenue as of Q1 2026, driven by training compute costs and the expense of serving a rapidly growing user base. The company is generating approximately $2 billion in revenue per month but spending well above that. Public investors will face the question of whether a $1 trillion valuation is justified by a path to profitability — or whether it prices in a platform-scale future that has not yet arrived. Anthropic is separately targeting an October 2026 IPO, per Bloomberg, and has not yet filed its own S-1. If both companies list in Q4 2026, it will be the first time two frontier AI labs have become publicly traded within the same quarter.

3. Anthropic and the Gates Foundation Commit $200 Million to AI for Global Health and Education

Anthropic and the Bill and Melinda Gates Foundation announced a $200 million, four-year partnership on May 14, 2026 to deploy Claude in global health, life sciences, education, and economic mobility programs. The commitment is split evenly: Anthropic contributes technical expertise and Claude usage credits; the Gates Foundation contributes grant funding and program design. The largest portion targets global health, particularly in low- and middle-income countries where approximately 4.6 billion people lack access to essential health services.

Specific programs include using Claude to accelerate vaccine candidate screening for diseases like polio and HPV before pre-clinical development; working with the Gates Foundation's Institute for Disease Modeling to improve disease forecasting for malaria and tuberculosis deployment decisions; and developing AI-powered K-12 tutoring tools for students in the US, sub-Saharan Africa, and India. The partnership will also build open-source African language datasets and release knowledge graphs to improve model performance in underserved language contexts — resources that will be publicly available across the industry, not proprietary to Anthropic or the Gates Foundation.

The deal is significantly larger than the Gates Foundation's previous AI partnership: a $50 million deal with OpenAI in January 2026 (Horizon1000) aimed at improving 1,000 primary healthcare clinics in Africa. Anthropic's $200 million commitment — four times that size — signals how aggressively both parties are treating this as a strategic priority, not a philanthropic side project. Anthropic is running this work through its Beneficial Deployments team, which already provides Claude credits and engineering support to nonprofit and education partners globally.