OpenAI burns $3.7bn in Q1 as revenue hits $5.7bn

🔑 Enhanced Key Takeaways

•OpenAI's 2025 financial results reported $13.07 billion in revenue against $34 billion in costs, leading to a $20.92 billion operating loss, with R&D spending accounting for 56% of total costs.
•The company projects spending $50 billion on computing in 2026, primarily for inference infrastructure to serve its 600 million weekly active users, training new models, and supporting research and safety initiatives.
•OpenAI successfully closed a $122 billion funding round in March 2026, achieving an $852 billion valuation, with significant contributions from Amazon ($50 billion), SoftBank ($30 billion), and Nvidia ($30 billion).
•To mitigate escalating costs and reduce reliance on third-party suppliers, OpenAI is collaborating with Broadcom to design its own custom AI chips, with deployment expected in late 2026 and mass production targeted for 2026 on a 3 nm process node.
•The enterprise segment now constitutes over 40% of OpenAI's revenue and is anticipated to reach parity with consumer revenue by the end of 2026, indicating a significant shift in its revenue streams.

📊 Competitor Analysis▸ Show

LLM API Pricing Comparison (as of May 2026)

Provider	Model	Input (per 1M tokens)	Output (per 1M tokens)	Context Window	Key Differentiators
OpenAI	GPT-5.5	$5.00	$30.00	N/A	Strongest model for complex reasoning, nuanced writing, agentic workflows.
	GPT-4.1	$2.00	$8.00	1,000,000	Ecosystem champion, default for standard production pipelines.
	GPT-4.1 Nano	$0.10	$0.40	1,000,000	Cheapest budget model, ideal for high-volume, cost-sensitive tasks.
Anthropic	Claude Opus 4.8	$5.00	$25.00	1,000,000	Elite coding and agentic performance, largest context window at flagship tier.
	Claude Sonnet 4.6	$3.00	$15.00	1,000,000	Offers annual billing savings on Pro, competitive for multimodal workloads.
	Claude Haiku 4.5	$1.00	$5.00	200,000	Higher cost than OpenAI's budget models.
Google	Gemini 3.1 Pro	$2.00	$12.00	1,000,000	Highly competitive for multimodal workloads (images, audio, video).
	Gemini 2.5 Flash	$0.30	$2.50	1,000,000	Clear cost leader in mid-tier, 10x cheaper on input, 4-6x cheaper on output than competitors.
	Gemini 2.5 Flash Lite	$0.10	$0.40	1,000,000	Tied with GPT-4.1 Nano as cheapest budget model.
xAI	Grok 4.3	$1.25	$2.50	1,000,000	Cheapest output at flagship tier, roughly 80% cheaper than Gemini 3.1 Pro.
	Grok 4.1 Fast	$0.20	$0.50	2,000,000	Offers an unprecedented 2M token context window for a budget model.

🛠️ Technical Deep Dive

OpenAI is actively developing its own custom AI chips, referred to as "AI accelerators," in collaboration with Broadcom. This initiative aims to integrate insights from frontier model development directly into hardware, enhancing performance and reducing dependency on external suppliers.
These custom chips are slated for mass production in 2026, utilizing TSMC's 3 nm process node.
The "Stargate Project," a joint venture with Oracle, SoftBank, and MGX, is a massive AI infrastructure undertaking estimated to cost $500 billion over four years, with an immediate deployment of $100 billion.
The Stargate project aims to secure 10 gigawatts of AI computing capacity, with initial sites under development in Texas, New Mexico, Wisconsin, and Michigan.
OpenAI's compute infrastructure relies on high-performance GPUs, including NVIDIA GPU clusters, and can scale to tens of millions of CPUs to handle demanding generative AI workloads.
Training frontier large language models (LLMs) like GPT-4 has been estimated to cost between $78 million and $100 million in compute alone, while Gemini Ultra 1.0 reached approximately $192 million, with compute infrastructure typically representing 60-70% of these expenses.

🔮 Future ImplicationsAI analysis grounded in cited sources

The high capital requirements for AI development will accelerate industry consolidation.

The immense costs associated with training and deploying frontier AI models, coupled with the need for vast computing infrastructure, favor large incumbents with substantial financial backing, leading to smaller players being acquired or exiting the market.

OpenAI's shift to custom AI chip design will increase its long-term control over performance and costs.

By developing in-house hardware, OpenAI can optimize chips specifically for its AI workloads, reduce reliance on third-party suppliers, and gain greater control over its supply chain and operational expenses.

AI investment will increasingly focus on tangible business value and multi-agent systems.

After a period of experimentation, businesses are shifting from AI pilot projects to strategic implementations that deliver measurable ROI, with a growing emphasis on multi-agent orchestration for complex workflows.

⏳ Timeline

2015-12

OpenAI founded as a non-profit organization.

2019-07

Microsoft makes an initial $1 billion investment in OpenAI.

2020-06

OpenAI announces GPT-3, a groundbreaking language model.

2023-01

Microsoft announces a multi-year, $10 billion investment in OpenAI.

2025-01

OpenAI launches the 'Stargate Project' with partners, committing $500 billion to AI infrastructure development.

2025-10

OpenAI partners with Broadcom to design its own custom AI chips.

2026-03

OpenAI closes a $122 billion funding round at an $852 billion valuation.

2026-04

OpenAI introduces GPT-5.5, its 'smartest model yet'.

OpenAI burns $3.7bn in Q1 as revenue hits $5.7bn

⚡ 30-Second TL;DR

🧠 Deep Insight

🔑 Enhanced Key Takeaways

LLM API Pricing Comparison (as of May 2026)

🛠️ Technical Deep Dive

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline

📎 Sources (25)

👉Related Updates

Taiwan debates NT$240bn investment in drone defense systems