Rigetti Computing (RGTI) Q3 2026 earnings review
Rigetti hits its chip accuracy goal, but sales drop
Rigetti Computing, a maker of superconducting quantum computers, hit its main engineering goal while its sales paused between single-machine deliveries. Chief executive Dr. Subodh Kulkarni said its 108-qubit machine reached 99.5% accuracy and signed a $100 million research agreement with the U.S. Department of Commerce. However, third-quarter sales fell 28% from a year ago to $1.4 million, and by our math the $20 million full-year plan leaves fourth-quarter revenue outside one Indian system down 65%.
| Revenue | $1.4M -28% from a year ago |
|---|---|
| 108-qubit system accuracy | 99.5% up from 99.1% last quarter, meeting the 2026 goal |
| Free cash flow | -$43.3M in Q3 by our math, on $26.2M of equipment spending |
| Full-year 2026 revenue outlook | ~$20M implies ~$9.1M in Q4, led by the $8.4M India order |
⚖️ Verdict: ⚪ Neutral
The story is unchanged because solving the near-term accuracy stall came alongside a slower long-term schedule and heavier cash burn. Reaching 99.5% accuracy after two flat quarters validates the multi-chip design, but the new mid-2028 target for 99.7% accuracy on 300 qubits pushes out the timeline set a year ago while quarterly free cash burn reached $43 million by our math.
The question now is whether hitting 99.5% accuracy and winning Commerce backing will bring repeat machine orders or leave sales dependent on one-off government projects. If commercial and national-lab buyers sign new contracts, hardware revenue will hold up after the fourth quarter; if not, sales will drop back after the single Indian delivery is booked. Fourth-quarter acceptance of that system and any newly priced 2027 orders will settle it.
🐂 Bull Case
The 108-Qubit Machine Hit Its Accuracy Goal
Quantum computers make errors when their basic units, called qubits, work together in pairs. Two-qubit gate fidelity measures how often a pair of qubits performs a step without an error, and Rigetti's flagship 108-qubit machine, Cepheus-1-108Q, had been stuck at 99.1% for two quarters.
"Our Cepheus-1-108Q system achieved our goal of 99.5% median two-qubit gate fidelity, demonstrating continued performance improvement at the 100-plus-qubit level and further validating our chiplet-based approach to scaling superconducting quantum computers," chief executive Dr. Subodh Kulkarni said. That jump cut the median error rate in half, from 0.9% to 0.5%, across twelve connected 9-qubit chiplets, and smaller sections of 80 and 50 qubits reached 99.6% and 99.7%.
What to watch: customer acceptance of the 108-qubit system at India's C-DAC computing center in the fourth quarter. Passing that live test will confirm the accuracy gain works outside Rigetti's own lab and release the order into revenue.
Washington Signed the $100 Million Research Deal
Rigetti has sought U.S. government funding to help pay for its next generation of hardware, though the deal requires giving away new shares while the quantum group already sits on nearly half a billion dollars in cash. In September, Kulkarni said, "We also signed a definitive $100 million agreement with the U.S. Department of Commerce" under the CHIPS Act to fund smaller control electronics, larger cooling systems and higher-connectivity chip manufacturing.
As a condition of the award, the Department took "a minority, non-controlling equity stake in Rigetti". By our math from the balance sheet, common shares outstanding rose by 9.0 million in the third quarter to 342.7 million, about 2.7% dilution despite little cash coming in from option or warrant exercises.
What to watch: the quarterly filing's note on the Commerce transaction. It will show the exact share price given to the government and the milestones required to unlock the $100 million in cash.
🐻 Bear Case
Revenue Dropped Between Single-Machine Deliveries
Sales of on-premises quantum computers drove a jump in revenue in the first half of 2026, making hardware deliveries the core of the bull case for commercial demand. Without a major machine shipment in the third quarter, however, revenue fell back to a fraction of the prior quarter's level.
- Third-quarter revenue: $1.4 million, down 28% from a year ago and down 73% from the second quarter
- Full-year 2026 revenue outlook: about $20 million, which leaves $9.1 million for the fourth quarter by our math
- Fourth-quarter revenue outside the $8.4 million India order: about $0.7 million by our math, down 65% from a year ago
The full-year plan depends on "assuming C-DAC customer acceptance and associated revenue recognition in the fourth quarter". Because that single Indian government machine makes up almost the entire fourth-quarter target, underlying cloud and research revenue is shrinking rather than building a steady base.
What to watch: new purchase orders with disclosed dollar values in the fourth-quarter report. Without new hardware contracts to replace the India delivery, revenue will drop again in early 2027.
Rigetti Pushed Out Its Next Chip Targets
A year ago, Rigetti's roadmap called for a 150-plus-qubit system at 99.7% accuracy by the end of 2026 and a 1,000-plus-qubit system at 99.8% by the end of 2027. After those dates went unmentioned last quarter, Kulkarni set a new intermediate goal to "demonstrate a 300-plus-qubit system with approximately 99.7% median two-qubit gate fidelity and gate speeds below 60 nanoseconds by mid-2028".
The company also kept its target for a 1,000-qubit system at 99.9% accuracy "within approximately three years", which now points to late 2029. While Rigetti frames the third quarter as steady roadmap execution, the new schedule pushes the 99.7% accuracy mark out by eighteen months and the 1,000-qubit mark out by nearly two years compared with its 2025 targets.
What to watch: the qubit count per chiplet when Rigetti details the 300-qubit design. Moving from 9-qubit chiplets to larger tiles on schedule is required to keep the mid-2028 target from slipping too.
Higher Research and Factory Bills Doubled Burn
Building larger cooling systems and chip-making gear is costing much more money each quarter. Operating loss widened 62% from a year ago to $33.2 million, and net loss excluding one-offs nearly doubled to $20.5 million.
- Research and development (the cost of designing chips and systems): $25.5 million, up 70% from a year ago and up 1,035 percentage points of revenue
- Selling, general and administrative (overhead and sales costs): $8.5 million, up 44% from a year ago and up 301 percentage points of revenue
- Equipment purchases: $26.2 million in the third quarter by our math, more than the $16.4 million spent in the entire first half
Because operating cash burn and equipment spending both stepped up, free cash flow reached negative $43.3 million in the third quarter by our math, nearly matching the negative $48.4 million of the first six months combined.
What to watch: fourth-quarter equipment purchases and government reimbursements. They will show whether the Commerce award offsets the new capital spending or if quarterly cash burn stays near $43 million.
Risks this quarter didn't answer
Three standing concerns around competition and the order book received no new data in the earnings release, although Rigetti again highlighted gate speeds "which is about 10,000 times faster than trapped-ion systems and 100 times faster than neutral-atom systems".
- DARPA Quantum Benchmarking Initiative Phase B: still unmentioned a year after Rigetti was left out of the initial selection; winning entry into Phase B would validate the new 99.5% fidelity with defense buyers.
- Commercial and supercomputing order pipeline: no new Novera or large-system orders were announced; a disclosed contract value for 2027 delivery would ease the post-C-DAC revenue drop.
- Remaining performance obligations and customer concentration: not in the release; the quarterly filing's footnotes will show whether contracted backlog grew above last quarter's $2.7 million.
💲 Other KPIs
Cash, cash equivalents and available-for-sale investments fell by $44.2 million from $541.3 million in the second quarter and $589.8 million at year-end 2025. The balance has declined for three straight quarters as Rigetti funds operations and dilution refrigerators without raising new cash from stock sales, though the quantum group still carries zero debt.
Total deferred revenue — cash customers have paid upfront for systems or services not yet delivered — edged down by $84 thousand from $4.01 million in the second quarter, up from $0.70 million a year ago. The balance is holding steady while Rigetti works toward customer acceptance on the prepaid C-DAC system in the fourth quarter.
Gross margin — the share of revenue left after the direct cost of delivering systems and contracts — rose 40 percentage points from 20.8% a year ago and 18.2 points from 42.6% in the second quarter. It has risen for four straight quarters, helped in the third quarter by a smaller mix of hardware installation costs.
🔮 Guidance
Unchanged in driver, newly quantified as a full-year total. The plan calls for about $20 million in 2026 sales, up 182% from $7.1 million in 2025. By our math, that leaves $9.1 million for the fourth quarter, up 385% from $1.9 million a year ago. The jump depends almost entirely on recognizing the $8.4 million C-DAC machine in India.
Re-dated and rescaled. Rigetti set a new milestone for a 300-plus-qubit system with about 99.7% median two-qubit accuracy and gate speeds below 60 nanoseconds by mid-2028. Earlier plans had targeted 99.7% accuracy on a 150-plus-qubit system by the end of 2026, so the accuracy timeline moved out by roughly eighteen months while the target size doubled.
Unchanged. The long-term plan still calls for roughly 1,000 qubits inside a single cooling fridge at 99.9% accuracy and gate speeds below 50 nanoseconds within about three years. Because the rolling three-year window moves forward each quarter, the implied completion date now sits around late 2029 rather than the end-of-2027 date given in 2025.
❓ Key Questions
What is the 2027 order backlog after the India delivery?
By our math, the $20 million outlook leaves only $0.7 million of fourth-quarter revenue outside the $8.4 million C-DAC system. Investors need the dollar value of the 2027 Pittsburgh Supercomputing Center order and any other signed hardware contracts.
How many shares and what milestones tie to the $100M award?
Common shares rose by 9.0 million in the third quarter as the Commerce Department took an equity stake. Management has not yet disclosed in the release how much of the $100 million was funded upfront or what milestones unlock the rest.
Why did third-quarter equipment purchases jump to $26.2 million?
By our math, third-quarter property and equipment spending of $26.2 million exceeded the $16.4 million spent in the first half. Management should clarify how much went to dilution refrigerators and whether quarterly capital spending will stay at this level.
Why does contracted backlog not match announced government contracts?
Last quarter's filing showed $2.7 million of remaining performance obligations against the $8.4 million C-DAC order and a $5.8 million Air Force contract. Until the third-quarter filing arrives, the true non-cancellable backlog remains unreadable.