Quantum squeezed states of light can enhance measurement sensitivity beyond classical limits and enable quantum information processing, but scalable low-loss sources remain challenging. We demonstrate continuous-wave quantum squeezing on a chip, achieving 18 dB of squeezing and 20 dB of anti-squeezing at 1570 nm in a 1.6-cm traveling-wave adaptively poled thin-film lithium niobate waveguide. A distributed model independently determines facet losses, phase noise, and nonlinear interaction strength without prior assumptions, enabling rigorous inference of on-chip performance. We estimate a 95

High-performance computing underpins modern artificial intelligence (AI), enabling foundation models, real-time inference and perception in autonomous systems, and data-intensive scientific simulations. Recent advances in quantization techniques utilizing low-precision computation without degrading model accuracy, create new opportunities for analog photonic computing characterized by ultra-high clock rates and low energy consumption. Here we propose and demonstrate a coherent homodyne integrated circuit capable of general matrix multiplication (GEMM) with aggregate throughput that exceeds 1,000 TOPS (tera-operations per second), enabled by massive on-chip optical fanout and parallelism. By leveraging time multiplexing, the required modulator count is reduced from O(N^2) to O(N), allowing dense integration of record-scale 256 × 256 homodyne units (each <0.0064 mm^2) within a single reticle. We employ wafer-scale fabricated 64 thin-film lithium niobate (TFLN) transmitters (each over 40-GHz bandwidth with propagation loss of 0.2 dB/cm) to encode data and chip-to-chip coupled to Si/SiN computing circuits (64 channels). Our system achieves up to 7-bit computational accuracy across 8 × 8 parallel channels at record computing clockrate 120 Gbaud/s, and 6-bit statistical accuracy across 256 × 100 channels at 20-128 Gbaud/s, representing a total throughput of 1,000-6,000 TOPS. Massive parallelism amortizes the optoelectronic (OE) conversion to allow 330-TOPS/W efficiency using foundry-available packaging technology. The system throughput is benchmarked with Qwen2.5-0.5 billion parameter models that generate accurate tokens. High throughput and energy efficiency establish a near-term pathway toward light-based accelerators for large-scale training and low-latency inference from datacenters to edges, accelerating new models toward artificial general intelligence.

Optical neural network processors offering benefits in bandwidth and energy consumption but problems in scaling and parallelism. We present a novel optical tensor processor capable of optically performing large-scale, highspeed matrix-matrix multiplication in a single step.

Although thin-film lithium niobate (TFLN) facilitates efficient signal generation and nonlinear and quantum interactions, the realization of optical parametric amplification (OPA) that can provide simultaneous ultra-broadband and high-gain operation in an integrated chip continues to pose a challenge. Here we demonstrate continuous-wave-pumped OPA in an X-cut MgO-doped, dispersion-engineered, and adaptively-poled TFLN waveguide of 1.6 cm length, achieving a flat-top profile covering a 450 nm-wide optical wavelength window, corresponding to a 3-dB gain bandwidth of about 56 THz. Among reported TFLN platforms, our device exhibits the broadest 3-dB gain bandwidth. The same device can be pumped either directly at visible wavelengths through second-order x(2) interactions or in the telecom band through cascaded x(2) processes, the latter eliminating the need for a high-power visible pump laser. We achieve maximum gains of 8.87 +- 0.39 dB and 10.79 +- 0.43 dB at on-chip pump powers of 74 mW and 170 mW for the direct and cascaded schemes, respectively. We further directly probe OPA gain across the 1650-1900 nm wavelength range, where experimental gain measurements have remained scarce. With a normalized on-chip gain of 1.21 dB per W per mm, our device sets a new benchmark among reported cascaded x(2) nonlinear processes. This work advances the realization of integrated optical parametric amplifiers, offering high efficiency, robust gain, ultra-broadband bandwidth, and continuous-wave operation, thereby enabling new capabilities for next-generation quantum sensing and photonic systems.

The mid-infrared (MIR) spectral regime is central to applications including remote sensing, precision spectroscopy, higher harmonic generation, and free-space optical communication. However, coherent and broadband MIR modulation remains challenging owing to high optical loss, limited bandwidth, and large drive voltages in existing platforms. Here, we overcome these challenges by deploying a suspended thin-film lithium-niobate (TFLN) based electro-optic (EO) platform co-designed with high-performance traveling-wave microwave (MW) electrodes. We demonstrate a record-low Vpi,DC of 2.3 to 4.3 V over a broadband MIR bandwidth from 2.4 to 3.6 um, and a 2.7 dB EO bandwidth of 40 GHz (extracted 3 dB bandwidth of 50 GHz), yielding a figure of merit of 17.4 GHz/V, more than an order of magnitude higher than the state of the art. We demonstrate, for the first time, high-frequency Vpi,MW of 4.5 to 6.5 V in the 25 to 35 GHz range, and frequency-agile MIR EO frequency comb generation with a 10 dB optical bandwidth over 0.8 THz using a suspended phase modulator of 4 cm active modulation length. We further validate the platform in a free-space optical communication link. Our results establish a monolithic MIR photonic platform capable of powerful EO modulation and spectral synthesis, and represent a significant step toward reconfigurable MIR sensing and communication systems on chip.

The mid-infrared (MIR) spectral region is central to sensing, spectroscopy, and free-space optical communication, yet coherent and broadband electro-optic (EO) control remains limited by material loss, bandwidth constraints, and high switching voltages. Existing approaches based on quantum cascade laser modulation, nonlinear frequency conversion, or bulk EO devices suffer from fundamental trade-offs between efficiency, bandwidth, and scalability. Here we report a suspended thin-film lithium niobate (TFLN) MIR EO platform co-designed with velocity- and impedance-matched traveling-wave microwave electrodes. We achieve record-low half-wave voltages of 2.3–4.3 V across 2.4–3.6 μm and EO bandwidths up to 50 GHz, corresponding to a record voltage-bandwidth figure-of-merit of 17.4 GHz/V. High-frequency operation with 4.5–6.5 V is demonstrated at 25–35 GHz, together with 30-GHz-line-spacing MIR EO frequency comb generation spanning 0.8 THz. We further validate the platform in a free-space communication link, establishing a scalable high-performance MIR optoelectronic platform.

Fundamental phase noise in thin-film lithium niobate (TFLN) photonic integrated circuits is governed by thermal-charge-carrier-refractive (TCCR) dynamics arising from thermally driven carrier fluctuations. In contrast to the predominantly thermorefractive noise in silicon photonic platforms, TCCR noise represents a distinct mechanism that becomes critical for applications requiring high frequency stability and phase coherence, including optomechanical sensing, low-phase-noise microwave synthesis, and on-chip quantum squeezing. A quantitative understanding of the deterministic parameters that control TCCR noise is therefore essential for engineering the next generation of low-noise TFLN photonic systems. Here, we identify two dominant contributors to the TCCR noise in TFLN microresonators: material anisotropy and surface states. Material anisotropy results in increased noise for extraordinarily polarized optical modes and leads to a geometry dependent phase noise. Surface-state effects manifest as increased noise in higher-order transverse modes as well as more than 120-fold higher noise in suspended microresonators. Finally, we demonstrate that post-fabrication annealing – widely used to reduce defect densities and recover crystal quality – suppresses frequency noise by a factor of 8.2 in cladded microresonators. Together, these results establish a practical pathway for noise engineering in TFLN integrated photonic devices and accelerate their deployment in next-generation precision photonic systems.

Modern problems in high-performance computing, ranging from training and inferencing deep learning models in computer vision and language models to simulating complex physical systems with nonlinearly-coupled equations, require exponential growth of computational resources. Photonic analog systems are emerging with solutions of intrinsic parallelism, high bandwidth, and low propagation loss. However, their application has been hindered by the low analog accuracy due to the electro-optic distortion, material nonlinearities, and signal-to-noise ratios. Here we overcome this barrier with a quantization-aware digital-photonic mixed-precision framework across chiplets for accelerated AI processing and physical simulation. Using Lithium Niobate photonics with channel equalization techniques, we demonstrate linear multiplication (9-bit amplitude-phase decoupling) in homodyne optical logics with 6-bit precision at the clock rate of 128 giga-symbol-per-second (128 GS/s), enabling AI processing with 6 ns latency. Codesign hardware-algorithms, including iterative solvers, sparse-dense quantization, and bit-sliced matrix multiplication, explore photonic amplitude and phase coherence for complex-valued, physics-inspired computation. In electromagnetic problems, our approach yields 12-bit solutions for partial differential equations (PDEs) in scattering problems that would conventionally require up to 32-bit and often even 64-bit precision. These results preserve digital-level fidelity while leveraging the high-speed low-energy photonic hardware, establishing a pathway toward general-purpose optical acceleration for generative artificial intelligence, real-time robotics, and accurate simulation for climate challenges and biological discoveries.

We realize an integrated electro-optic Fresnel time lens on a thin film lithium niobate (TFLN) platform, achieving a coherent interface between 1.1 nanosecond and 16 picosecond pulses using 143 mW of microwave (MW) power.

The ever-increasing data demand craves advancements in high-speed and energy-efficient computing hardware. Analog optical neural network (ONN) processors have emerged as a promising solution, offering benefits in bandwidth and energy consumption. However, existing ONN processors exhibit limited computational parallelism, and while certain architectures achieve high parallelism, they encounter serious scaling up roadblocks for large-scale implementation. Here, we introduce a spatial-wavelength-temporal hyper-multiplexed ONN processor, which is based on parallel diffractive beam routing. The architecture supports high three-dimensional data, high O(N3) computing parallelism, and is feasible for large-scale implementation. A 16 × 16 parallel diffractive beam routing is demonstrated, enabling a large-scale (16 × 16 - by - 16 × 16), high-parallelism (4096 multiply-and-accumulates/shot (MACs/shot)), high-speed (2 Gsa/s), single-shot matrix-matrix multiplication (MMM) optical tensor processor. It accelerates convolutional neural networks (CNNs) and deep neural networks (DNNs) through parallel matrix multiplication. We demonstrate benchmark image recognition using a CNN and a subsequently fully connected DNN in the optical domain. The network works with an ultra-low optical energy of  ≈ 20 attojoules (aJ)/MAC at 96.4% classification accuracy. The ONN system supports broad spectral and spatial bandwidths and is capable for large-scale scaling up, paving the way for highly efficient large-scale optical computing for next-generation deep learning.

Optomechanical transduction harnesses the interaction between optical fields and mechanical motion to achieve sensitive measurement of weak mechanical quantities with inherently low noise. Lithium niobate combines low optical loss, strong piezoelectricity, high intrinsic fQ_m factor, and low thermal conductivity, making it promising for exploring optomechanical platforms targeting thermal sensing applications. Here, we developed an integrated optomechanical platform on thin-film lithium niobate with precisely engineered optical, mechanical, and thermal fields within a compact 40 μm by 40 μm footprint. The platform integrates suspended microring resonators with ultrathin central membranes, reducing mechanical stiffness and effective mass while maintaining a high optical factor Q_o of 1e6 and mechanical quality factor Q_m of 1117, which increases to 5.1e4 after oscillation. The design suppresses thermal dissipation into the silicon substrate and enhances thermal sensitivity, achieving a temperature coefficient of frequency of -124 ppm/K and a noise-equivalent power of 6.2 nW/sqrt(Hz) at 10 kHz at room temperature. This compact and scalable platform opens up new opportunities for high-sensitivity thermal sensing, supports heterogeneous integration with infrared absorbers for uncooled infrared detection, and enables fully integrated, all-optical on-chip readout, paving the way toward large-format, low-noise infrared sensing arrays.

The twin fields of ultrafast optics and nonlinear photonics enable applications ranging from attosecond science [1, 2] and ultrafast electronics [3] to molecular spectroscopy [4, 5], nonlinear optics [6-8], quantum nanophotonics [9] and precision metrology [10]. However, bringing these capabilities-including ultrashort pulse generation, dispersion control, and strong nonlinear interactions-together within a scalable photonic integrated platform requires exceptional performance and cooperation between components while preserving sufficient optical power across the circuit. Here we demonstrate an integrated multi-functional nonlinear photonic system that transforms continuous-wave (CW) light into high-peak-power femtosecond pulses and harnesses them for pulse-driven nonlinear optics on thin-film lithium niobate (TFLN). Microwave-driven electro-optic (EO) broadening followed by integrated dispersive compression generates 230-fs Fourier-transform-limited pulses with energies up to 3.3 pJ at 30.7 GHz, representing orders of magnitude higher pulse energy than previous integrated pulse synthesis at comparable repetition rates [11]. In a 0.3-meter dispersion-engineered TFLN waveguide, soliton dynamics compress the pulses to 35 fs (6.7 optical cycles), accompanied by coherent spectral broadening exceeding 330 nm. In a fully-monolithic architecture, EO synthesis, dispersive compression and a high-Q nonlinear resonator are integrated on a single TFLN chip, enabling resonantly-enhanced pulse pumping and coherent spectral broadening at pulse energies as low as 400 fJ. By unifying microwave-controlled pulse synthesis and pulse-driven nonlinear interactions, our work establishes a direct path from CW excitation to few-cycle nonlinear optics on chip, with opportunities spanning microwave photonics [12], optical frequency synthesis and metrology, and mid-infrared and terahertz generation [13, 14].

Quantum noise limits the sensitivity of optical measurements, but squeezed states of light enable quantum-enhanced metrology, sensing, and information processing. Most on-chip squeezed-light sources rely on Kerr (χ^(3)) nonlinearities, remain limited by pump power and excess loss constraints. Quadratic (χ^(2)) platforms instead provide stronger parametric interactions, lower pump power requirements, and greater spectral engineering flexibility. Here, we demonstrate strong, broadband squeezed-light generation on a thin-film lithium niobate (TFLN) photonic chip using a dual-resonant optical parametric amplifier implemented in a single periodically poled LN (PPLN) microresonator. Near-full-depth domain inversion is achieved simultaneously with highly over-coupled resonances, exhibiting escape efficiencies exceeding 90

The ability to process and act on data in real time is increasingly critical for applications ranging from autonomous vehicles, three-dimensional environmental sensing, and remote robotics. However, the deployment of deep neural networks (DNNs) in edge devices is hindered by the lack of energy-efficient scalable computing hardware. Here, we introduce a fanout spatial time-of-flight optical neural network (FAST-ONN) that calculates billions of convolutions per second with ultralow latency and power consumption. This is enabled by the combination of high-speed dense arrays of vertical-cavity surface-emitting lasers (VCSELs) for input modulation with spatial light modulators of high pixel counts for in-memory weighting. In a three-dimensional optical system, parallel differential readout allows signed weight values for accurate inference in a single shot. The performance is benchmarked with feature extraction in You-Only-Look-Once (YOLO) for convolution at 100 million frames per second (MFPS), and in-system backward propagation training with photonic reprogrammability. The VCSEL transmitters are implementable in any free-space optical computing systems to improve the clockrate to over gigahertz, where the high scalability in device counts and channel parallelism enables a new avenue to scale up free space computing hardware.

We demonstrate an analogue frequency-comb-based optical smart sensor with in-memory computing for real-time image classification and blood analysis. Our system successfully predicts early-stage heart-attack samples with 95.1% accuracy with single-pulse latency of 10ns.

The exponential growth in deep learning models is challenging the fundamentals of digital electronics and motivating new computing paradigm. Optical neural networks are emerging to accelerate machine learning tasks with high optical bandwidth, parallelism, and low-loss data movement. However, the scalability of existing optical accelerators is limited by the low electro-optic conversion efficiency, large photonic device footprints, lack of optical nonlinearity, etc. In this talk, I will present our computing architectures based-on space-time or wavelength-time multiplexing techniques to overcome the electronic-optic-electronic bottlenecks and scale up optical neural network systems. We explore different photonic platforms, including microlaser arrays, silicon photonics and thin-film lithium niobate photonics. Our demonstration has experimentally realized a full-system energy efficiency of 140 TOPS/W and a chip computing density of 6 TOP/(mm2·s), representing 100-fold improvement over state-of-the-art digital processors.

The stability of the integrated photonic circuits is of critical importance for many applications that require high frequency precision or robust operation over time, such as optomechanical sensing, frequency conversion, optical communication, and quantum optics. Photonic memory is useful for low-energy optical computing and interconnects. Thin film lithium niobate (TFLN), as an emerging photonic platform, exhibits complex material properties including pyroelectric (PE) and photorefractive (PR) effects which could lead to intra-device drift and excess noise under different environmental or operating conditions as well as be utilized for building photonic memory. However, the long-term stability and memory effect of its optical properties has not been explored. In this paper, we discovered a long-lived change of optical refractive index as a result of light excitation and temporal temperature variation using Z-cut TFLN microresonators and reveal a strong dependence of the instability with the crystal orientation of the thin film form. The recovery time are measured to be over 10 hours. Leveraging the photonic memory with a long relaxation time, we realize optical trimming of the cavity resonance frequencies. Our result offers insights towards understanding the fundamental noise properties and dynamic behavior of the integrated TFLN material and devices.

We demonstrate large-scale VCSEL-based computing for convolution neural networks with spatial fanout and high clockrates. We achieved 9 parallel channels with >5 bits precision at 100-million shot-per-second, with in-system training over 93.4% accuracy.

Edholm's law predicts exponential growth in data rate and spectrum bandwidth for communications. Owing to exponentially increasing deep neural network computing demands and the slowing of Moore's law, new computing paradigms are required for future advanced communications like 6G. Optical neural networks (ONNs) are promising accelerators but struggle with scalability and system overhead. Here, we introduce our multiplicative analog frequency transform optical neural network (MAFT-ONN), an artificial intelligence hardware accelerator that experimentally computes fully analog deep learning on raw radio frequency (RF) signals, performing modulation classification that quickly converges to 95% accuracy. MAFT-ONN also exhibits scalability with nearly 4 million fully analog operations for MNIST digit classification. Because of the Shannon capacity-limited analog data movement, MAFT-ONN is also hundreds of times faster than traditional RF receivers.

We demonstrate an air-suspended thin-film-lithium-niobate Mach-Zehnder intensity modulator for mid-infrared communications. Data channels of various modulation formats are achieved, including 1.5-Gbaud OOK, 1.5-Gbaud PAM-4, 0.5-Gbaud PAM-6, and 0.5-Gbaud PAM-8.